Computer Integrated Steel Bridge Design and Construction
Expanding Automation
Final Report

WORKSHOP FORMAT

General Session

In preparation for the workshop, a Pre-Workshop Survey and Questionnaire and a Pre-Workshop Fabricator Questionnaire were sent to prospective workshop attendees. The questionnaires asked the attendees to comment on the major obstacles toward automation in our industry. The information contained in their replies helped workshop organizers structure the general session and breakout sessions of the workshop.

The workshop's first day was devoted to hearing presentations regarding automation and computer integrated systems from domestic and foreign sources. The second and third days were spent brainstorming ideas and writing action plan statements. First day presentations included the following and involved authors from Europe and Japan as well as the U.S. (2):

• Observations from FHWA's Steel Bridge Automation-International Review Tour (K. Frank, U. of Texas)
• Looking for Optimal Factory Automation Balanced Against Demand Capital (S. Tada, Kawada Industries)
• Electronic Data Transfer and Processing for Automation (S. Chen, SUNY at Buffalo)
• Automated Fabrication and Detailing Methods in the UK (G. Booth, Fairfield Mabey)
• Bulldozers or Bridges--What's the Difference (T. Jutla, Caterpillar, Inc.)
• Automation and Process Improvement in the Shipbuilding Industry (J. Dydo, Edison Welding Institute)
• Systemization with CAD/CAM Welding Robots in Steel Bridge Fabrication (Y. Kanjo, NKK Industries)
• Advances in Arc Welding Technology for Heavy Manufacturing (D. Harwig, Edison Welding Institute)
• Steeling the Competitive Edge - Is there a place for Robots (J. Weston, The Welding Institute)
• Automation Generation of Contract Plans for Steel Bridges (J. Jang and R. Teli, Columbus Engineering Consultants)
• Computer Integration in Fabrication and Erection (J. O'Neil, Cleveland Bridge)

Opening Session

Welcoming Remarks

Mr. Krishna Verma, Senior Welding Engineer with the Office of Bridge Technology in the Federal Highway Administration (FHWA), welcomed the assembled participants on behalf of the sponsoring agencies and pointed out that the 1999 International Review Tour served as the basis of this workshop. That tour revealed to U.S. experts that steel bridge members can be fabricated efficiently and economically using automation and robots. Still, no fully integrated design/fabrication/erection process exists currently, as steel bridge fabrication practice in the U.S. is currently highly decentralized. With the large number of steel bridges that are fabricated in the United States each year, and with our expanding bridge program, Computer Integrated Manufacturing technology has tremendous potential.

Mr. Arun Shirole', Executive Director of the National Steel Bridge Alliance (NSBA), welcomed the participants and pointed out the diverse constituencies they represent: the public sector (owners), fabricators, academics, software developers, and the Edison Welding Institute (EWI). This workshop is believed to be the first of its kind anywhere.

Plenary Session 1: Automation in Steel Bridge Design and Construction - Current Status

"Observations From FHWA's Steel Bridge Fabrication Review Tour" Prof. Karl Frank of the University of Texas at Austin, a participant in FHWA's 1999 International Review Tour, presented a summary of observations relating to aspects of automation from that tour of steel bridge fabrication plants in Japan, Italy, Germany, and the United Kingdom. Although none of the plants were fully automated, the ones with the highest degree of automation had fully automated thermal cutting lines, girder lines and web lines which included automated stiffener welding, and a semi-automated stiffened plate straightening station.

Implications for changes to U.S. practice, if these advances were to be deployed here, include the following:

• Elimination of radiographic inspection in favor of automation-friendly ultrasonic inspection, which would require new definitions of equipment and operator qualifications and new acceptance specifications based on fitness for purpose rather than the present workmanship requirements,
• Elimination of submerged-arc welding (and required flux handling systems) in favor of automation-friendly GMAW or MIG/MAG welding processes, and
• Use (and long-term archival) of a single 3D CAD model as the sole source of information on detailing, shop drawing information, CNC drilling and cutting instruction, automated inspection and virtual assembly (geometry verification), and
• Possible contractual ties between fabricator and erector in order to facilitate virtual assembly

"Looking for Optimal Factory Automation Balanced Against Demand, Capital Investment and Efficiency"

Mr. Satoshi Tada of Kawada Industries presented a description of the automation history of the Shikoku plant in Japan, where steel box girders are much more prevalent than in the U.S. The CAD/CAM system in this plan provides a 3D model of fabrication geometry. This model enables checking of all three-dimensional relationships of all components within the structure and virtual assembly. It also generates CNC fabrication data obtained through "a 2D unfolding process within the system."

Another feature of the fabrication plant is extensive use of robotic welding equipment, processes, and attendant automation-friendly detailing. Interestingly, there is little automation downstream (box assembly, welding, straightening, and finishing). Quantitative comparisons of costs to produce project documentation before and after automation were presented.

"Electronic Data Transfer and Processing for Automation"

Prof. Stuart Chen of State University of New York (SUNY) at Buffalo described the need for standardization and automation in the context of the currently fragmented nature of the industry, along with a review of some current developments towards standardization and a case study showcasing benefits of automation based on 3D CAD modeling. Standardization should focus on the specifications (intellectual requirements) that "stay the same" from job to job, e.g., design details, inspection methods and acceptance criteria. Automation is envisioned via design and detailing software, hardware and software for electronic information transfer, and fabrication equipment and processes. A review of the various consensus standards developed (or under development) by the AASHTO/NSBA Steel Bridge Collaboration highlights aspects of both the content of and the process underlying standards that owners/specifiers need to adopt before automation and its attendant economies can move forward. What these standards accomplish for information transfer for human eyes needs to be extended to electronic means of information transfer. Then the present "islands of automation" will begin to be bridged, likely based on rapidly developing web-based standards for software integration and business-to-business electronic commerce.

Benefits of 3D CAD modeling in a steel bridge rehabilitation project were highlighted in a case study of a deck replacement. The 3D modeling in combination with a GIS-based site survey was indispensable not only for fabrication but also for the visualization and planning of construction sequencing that involved installing a temporary gantry crane system, maintaining traffic, replacing utility lines and girders, jacking and re-using existing trusses, and pouring a new deck.

EWI Facilities Tour

Dr. Robert Kratzenberg and his associates of the Edison Welding Institute (EWI) hosted a tour of EWI's internationally recognized laboratory and training facilities.

Luncheon Keynote Session

"Automated Fabrication and Detailing Methods in the U.K."

Mr. Geoffrey Booth of Fairfield-Mabey in England described the use of fully integrated 3D modeling, which in spite of being 30% slower than 2D CAD has huge spin-off benefits. The single 3D model is the basis for CAM (computer-aided manufacturing) as well as CAD, since the model serves as a template for CNC layout. The model is used to inform machines, not shop floor people. Thus, there are no shop drawings (!). The model is a single source of data about all relevant attributes of the structure that can be packaged in various ways. Quantity takeoffs for estimating and scheduling, for example, can be extracted from the single database of structure attribute information.

Fabricator workflow involves issuing RFI's (requests for information) to designers upon scrutinizing the designer's drawings, with the 3D model generated after answers are received. Full numerical descriptions of each attribute of each piece enable generation of plate nesting instructions as well as CNC machine instructions. It is suggested that the 3D modeling system is the most significant advance and is the one to implement first; robotization can then follow. A "sensible mix of human skills and hi-tech machines" is desirable, "otherwise the technology owns you." Weld preheat does not fit well within automated processing and the need for it should be eliminated. ISO 9001 certification is used as the basis of quality assurance. The most typical contracting mechanism in the U.K. is DBFO (Design-Build, Finance, Operate), where the contractor recruits the designer and fast turnarounds are valued.

Plenary Session 2: Experience with Automation for Steel Fabrication

"Bulldozers or Bridges - What's the Difference?"

Dr. Tarsem Jutla of Caterpillar described the digital design and manufacturing employed by Caterpillar's $3.5 billion business in fabricated structures, specifically earth-moving machines that are more complicated than our bridge structures, with tighter tolerances. Computer simulations using high-powered finite element analyses extracted from CAD solid models are employed, for example, for the following:

• rolling and flattening operations, including cold rolled residual stress distributions,
• laser cutting, including thermal distortions.

Thus, adaptive forming can compensate for "springback" and welding process simulations predict residual stress and distortions and enable robot process planning to work right the first time.

This presentation essentially pointed out, via the illustrative example provided by Caterpillar bulldozers, that customized job-shop fabrication of large steel structures could be highly automated around robust 3D computer modeling and that such automation increased throughput and reduced turnaround time. Future trends point to increased use of laser technology not just for cutting (currently up to 35 mm thick) but also for shaping parts, removing mill scale and machining. Working closely with the supply chain is considered important to establish strategic relationships to share both risks and rewards.

"Automation and Process Improvements in the Shipbuilding Industry"

Dr. James Dydo of Edison Welding Institute provided an overview of current procedures and recent developments in steel fabrication within the shipbuilding industry. Increased use of robotics and laser process methods are evident, as are changes in design for fabrication, e.g., design for modular assembly. Robust techniques that reduce welding distortion and improve fitup accuracy are of interest to the Navy. Accordingly, they have initiated the following development projects (described further in the full paper):

• Advanced modeling techniques for prediction of distortion: buckling, angular change, and in-plane shrinkage ("We can predict distorted shape"),
• "Thermal tensioning" techniques that reduce the amount of welding distortion via application of auxiliary heat and/or cooling during the welding process, and
• "Thermal forming" techniques that apply heat selectively to the surface of a plate for the purposes of inducing curvature, even the compound (multi-axis) curvatures required of ship hulls.

"Systemization with CAD/CAM Welding Robots in Steel Bridge Fabrication"

Mr. Yoshihiro Kanjo of NKK Industries described multi-robot CAD/CAM welding systems developed for bridge fabrication and shipbuilding at NKK's Tsu works. The key technologies facilitating welding automation in the robotic systems are focused on four areas:

• High-speed rotating arc welding process for increasing welding efficiency,
• A coordinate transformation system tied in with multifunctional arc sensors that corrects the positional mismatch before and during welding,
• An altered production process (panel subassembly distinct from box assembly) in order to accommodate multiple simultaneous robots for welding for increased throughput, and
• An integrated 3D CAD/CAM computer system that generates not only 2D drawing views, material lists, and CNC data, but also robot motion simulation and path data (in conjunction with the welding motion pattern database) and the multi-robot control system (e.g., to avoid robot collisions during welding).

"Advances in Arc Welding Technology for Heavy Manufacturing"

Mr. Dennis Harwig of Edison Welding Institute presented emerging arc welding technologies and opportunities for heavy manufacturing. Principal drivers for these developments are the need for increased welding speed and higher quality weldments. Key technologies are identified as the following:

• high deposition processes (possible with SAW, FCAW, or GMAW-T),
• welding procedure optimization accounting for interactions between process and production factors,
• real-time data acquisition and quality monitoring with statistical process control,
• robotics using both cartesian and articulated arm approaches,
• adaptive welding technology with real-time through-arc sensing and dimensional inspection and application-specific adaptive fill algorithms, and
• hybrid welding (e.g., hybrid laser/GMAW welding).

Plenary Session 3: Implementation Issues

"Steeling the Competitive Edge - Is there a place for Robots?"

Mr. John Weston of the Welding Institute in England described the manufacturing processes used in the steel fabrication industry in terms of requirements for automation and robotics. Fabrication shops have been quick to implement islands of automation for operations such as CNC drilling, cutting, or machining. But there has been a reluctance to adopt a more widespread robotization of fabrication shop operations. These operations include:

• Cleaning (typ. blast cleaning), where a robotic arm could manipulate the blasting nozzle,
• Cutting and profiling, where robots in the automotive industry have been used in place of press trimming or punching,
• Hole forming, which can be done via drilling or punching on multi-headed NC machines in 3D on a range of section shapes,
• Joining, including not just arc welding robots but also bolting, nailing, riveting, and bonding with adhesives,
• Bending and pressing, where NC operation is increasingly being deployed,
• Rolling, where NC operational controls are increasingly being deployed,
• Machining, a longtime NC process where robots are increasingly being used,
• Applications of protective coatings, and
• Handling, which has thus far resisted robotic solutions owing to the sheer size and diversity of product dimensions and shapes.

Soon, robots may be applied in the following areas:

• Cutting and marking (increasingly with laser systems),
• Welding (initially using arc processes and eventually lasers),
• Nondestructive examination (NDE), and
• Coatings (painting robots), with their increased health and safety requirements.

Information technology (IT) is seen as a key enabling technology that in this context requires robust linkages between numerous computer based applications such as CAD, simulation cells, administration, scheduling, planning, purchasing (ordering), and maintenance control.

"Automated Generation of Contract Plans for Steel Bridges"

One of the disconnects in the process of producing a steel bridge is the need to generate contract drawings after the bridge is actually designed. This presentation focuses entirely on this particular disconnect. Dr. Jack Jang and Mr. Raju Teli of Columbus Engineering Consultants described commercial software, cecSTEEL, that generates contract plans compliant with Ohio standards, for steel beam and plate girder bridges. Superstructure drawings, for example, include:

• Typical transverse sections,
• Deck reinforcing plan,
• Framing plan and beam elevations,
• Deflection/camber table,
• Screed elevations table,
• Bearing and splice details, and
• Parapet transitions.

In effect, the development of the software required the parameterization of the information contained in a set of contract plans for an entire typical steel bridge crossing. Secondary calculations such as detailed structure dimensions, beam seat elevations, screed elevations, material quantities, pay item quantities, reinforcing steel list, etc., are directly performed by the software in order to reduce the tedium of data entry. The software prompts the user for only the governing design parameters, performs all secondary calculations, and then generates the AutoCAD or MicroStation files containing the contract plans for the bridge in a fraction of the time it would normally take.

"Computer Integration in Fabrication and Erection"

Mr. James O'Neill of Cleveland Bridge in England described construction methods employed in building two very different large bridge projects: asymmetric box girder bascule bridges on the Bellmouth crossing in London and the Boyne cable-stayed bridge in Ireland. Modularization was successfully employed to minimize site risks associated with working at heights and reducing construction period by maximizing off-site activities. Planning of construction, using purpose-built temporary decks and incremental launching assisted by temporary stays, was greatly facilitated by use of the 3D CAD model. This presentation highlights the role of automation in the fabrication - erection interface.

Plenary Session Summary

There are a number of error-prone, inefficient and time-consuming disconnects at the transitions between the various stages in the design, plan generation, fabrication, and erection of a steel bridge. The Day 1 introductory and plenary presentations at the workshop each examined various aspects of those disconnects and the emerging technologies that potentially could help to remedy them.

Some recurrent themes coming out of these presentations (and in the breakout sessions that followed) highlighted the need for the following:

• Solid modeling based on a single 3D model as in other industries (3), not just 2D CAD drafting, enabling direct links to manufacturing (e.g., CNC machines)
• Virtual assembly
• Model as a single source of data that can be extracted and packaged in various ways
• Automated inspection and data recording as being done elsewhere and in related industries,
• Use welding processes (e.g., GMAW) suitable for automation,
• Modifying/eliminating current specs that are unduly restrictive and outdated,
• Development of standard specifications, to replace the current myriad variety of state specifications, that would facilitate automation,
• Design-Build type of approach that overcomes the adversarial barriers that plague the traditional contracting approach typically employed in bridge construction.
• A cost effective systems approach that does not merely introduce robots or increase the speed of one of the processes.

The table below provides an overview of some of the recurrent themes, by speaker.

Plenary Speaker	3D CAD Model	Automated Inspection and Data Recording	Automation-Friendly Welding (e.g., SAW GMAW)	Automation-Friendly Inspection (e.g., RT UT)	Virtual Assembly	Specs: Need to Consolidate and Modernize	DB(FO)	Prediction &/or Control of Fabrication Operations	Systems Approach
Frank
Tada
Chen
Booth
Jutla
Dydo
Kanjo
Harwig
Weston
Jang
O'Neil

Implementation Panel Discussion and Open Forum

As a precursor to the small-group brainstorming sessions to follow, the entire group of 53 participants engaged in some collective brainstorming. Some of the opinions and suggestions expressed during this collective brainstorming, slightly edited, were as follows:

1. The model of information flow presented by Mr. Kanjo was helpful.
2. Focus first on 3D information modeling (rather than robotics).
3. We see that it can be done. What are the barriers to doing it here? We need to spend some time dealing with these barriers.
4. Design-build breaks down the adversarial barriers that our system has in it. The issue is who is taking the risk, and how to share the risk. Construction Managers (CM's) are not builders.
5. (In the U.K.) through DBFO, we're back to managing work, and expertise is valued. The issue is culture. When "FO" was added, expertise became valued.
6. Can we (in the U.S.) invest useful energy into DBFO? Perhaps DB advocates should be encouraged to push for DBFO.
7. We should define a list of items that need to be standardized along with associated impacts, e.g., gas welding and alternatives to RT (i.e., automation-friendly fabrication and inspection technologies).
8. Caterpillar, like the auto industry, produces customized, made-to-order products. What do they standardize?
9. Issue: how to set standards that make use of the knowledge out there (e.g., regarding welding). The basis for such standards ought to be fitness for service.
10. We need to articulate the benefits of change.
11. Comments thus far emphasize that lack of standardization is a hurdle. The steel industry keeps saying that they will give customers what they want, rather than articulate benefits of standardization.
12. DOTs at times have inexperienced engineers in charge of fabrication. They in turn are reluctant to consider spec changes.
13. Problem: how far can the owner be expected to bend?
14. A further problem is that owner-fabricator exchanges don't get communicated back to the designers.

Breakout Sessions

The objectives for the breakout sessions and group sessions on the second and third days were as follows:

1. Identify advances needed in the state-of-the-art of the various technical support technologies (e.g., robotics, open software standards), and
2. Identify high-payoff pilot projects, potential teams for those projects, and mutually agreed-upon statements of "where do we go from here?" to explore implementation issues and business process re-engineering required to implement available and emerging technologies needed.

On the second day, four breakout groups were formed based on the pre-workshop questionnaires (shown in Appendices B and C) and first day discussions. Each breakout group was constituted to have a mix of fabricators, DOT engineers, and other stakeholders. The breakout groups' first task was to brainstorm ideas on what the steel bridge industry's focus should be regarding automation. The four breakout groups identified over 100 ideas, listed in Appendix E. Participants then voted to identify the most important ideas for further development.

Workshop facilitators then organized these ideas into common themes. The breakout groups reconvened to write action plans for the resulting four theme areas. The action plans included objectives, rationale, short-, medium- and long-term tasks, potential resources to implement the plan, obstacles and potential payoffs. The four theme areas are as follows:

1. Computer Generated Drawings/Modeling and Electronic Information Transfer (Electronic Design and Drawings Transfer and Modeling)
2. Standardized Specifications and Paperless Approval Processes
3. Standardized Design Details to Facilitate Automation
4. Showcase of Benefits of Automation

The action plans for these four theme areas appear in Appendix A of this report.

Background | Table of Contents | Workshop Outcome